AnalyticBridge
A Data Science Central Community
Opinions on CCSU Masters in data mining?
I'd like opinions on Central Connecticut State University's Masters in data mining. I'm particularly curious to know if it is possible to get a decent job in data mining just by having this Master's, without an undergrad degree in a quantitative field such as comp sci, stats, or math. (I have a bachelor's in an unrelated field.)
Ideally I'd like to go back to school for the undergrad comp sci degree first, but it is looking prohibitively expensive to do so. I suppose I can pick up some programming and database skills on my own. It does seem like programming and database skills are pretty important in this field.
I want to be able to do interesting applied statistical analysis, data mining, and text mining for a medium-to-large company.
Also, are there comparable programs in the US? That is, affordable ones that will let me get a decent data mining job in two years or so?
Replies to This Discussion
-
Permalink Reply by Jeff on
-
Josh -
My 2 cents.
You need calc and linear algebra for the program, along with a first course in probability - so you need to have or acquire this background at least. You will get a lot of experience in data mining using IBM Modeler and R on real data sets and will learn linear models and multivariate analysis (cluster, factor MANOVA etc). There is also a course in math/stats to pick up basic concepts. There is also a text ming course that will teach you some Perl. So you can pick up a lot.
Many of the students work in the field so they do come with experience in IT, stats or other field, but not necassarily.
To be successful and be employed as a data miner (different than thinking any program will make you a data miner) you need to be curious and still do a lot of reading / research on your own (because you really love the subject). Programming languages are important to manipulate and extract data (R, SQL fro example) and you need a business / scientific background to be truly successful -- you have to frame data mining results in terms of ROI for the business side.
Hope that helps.
-
-
Thank you Jeff, I consider that to be a pretty encouraging answer. I do find data mining to be a fascinating subject, because I believe it to be the best way to understand how the world of commerce and other human affairs works, practically speaking. I didn't hear of it until a few years ago, but when I did I was immediately struck by the possibilities for interesting analysis it offers. And I do have a little background in IT / enterprise reporting, but it is all self-taught.
My current plans are to do the math and statistics prereqs at a community college, then go on to the CCSU program. I may take some business and programming classes at the community college as well. I can teach myself some programming. I'm messing around with R right now. Did you go through the CCSU program yourself?
I do have a question as to what you meant by "you need a business / scientific background to be truly successful". I understand that the business needs of course drive the whole project, and I have a little experience translating data for managers and execs. I consider the bridge between business and IT to be the sweet spot of employment for me personally. But I do not have any formal training in business, nor do I have a scientific degree or experience in a scientific role. I do have a scientific mindset though, I'm always looking for proof. I was thinking I could take some business classes as well to understand that side better. But I still question if there's a key ingredient I'll be missing by not having the scientific experience, and is there a way to make up for that?
-
-
By business / scientific background, I simply meant that you need to have domain knowledge in the field you are working (applying DM) to really be successful. For example, I work in CRM so for me that means I have to understand how to think in terms of translating findings into increased customer value (and it means thinking of the objective up front!). Or perhaps it means that when building a model and applying feature selection, knowing what the variables mean and using theory to guide the process along with the algorithms. I don't think I would take "business" courses per say, ultimately the more technical skills the better, but it would be worth understanding the basic business objectives for the domain(s) you wish to apply your DM trade to. Maybe seek out data mining articles and case studies around that domain. Good luck!
-
Permalink Reply by Eric Flores on
-
Josh,
I'm in the program, and in my opinion it is one of the best available because the focus of the program is very applied. I find it to be fascinating. I have seen other programs that have statistics courses that explain the theory behind data mining, but they do not seem to be so applied as this one.
However, I recommend you not to take our word for granted. Nobody will be able to answer questions related to your specific situation better than the program director, Prof. Larose. You can reach him at [email protected] . He is a very approachable person. You can tell him about your experience (or lack of), your skillset, your objectives and fears, everything.
One thing, though, is that the courses are very demanding. Actually, that can even be an understatement. They are not so intellectually hard or difficult to understand, but they require A LOT of work from you (reading texbook, learning to use the programs, processing data, writing reports, etc.). If you have family and a current full time job, dont ever consider taking two courses at the same time unless you have a strong statistics background and are experienced using R. Actually, having a robust statistics background helps a lot, even for those courses that have no formal statistics prerequisite. As Jeff says, linear algebra would help too, but I dont have it either. And about Calculus, I would recommend you to talk about it with Prof. Larose if you dont have it. Dont despair if you dont,
Regards,
Eric
© 2021 TechTarget, Inc.
Powered by
Badges | Report an Issue | Privacy Policy | Terms of Service
